research information letter 102, 'structural building ...seismic safety criteria for nuclear power...

:;- • • UNITED STATES

NUCLEAR REGULATORY COMMISSION

WASHINGTON, D. C. 20555

-

MEMORANDUM FOR: Harold R. Denton, Director

FROM:

SUBJECT:

Office of Nuclear Reactor Regulation

Robert B. Minogue, Director Office of Standards Development

Thomas E. Murley, Acting Director Office of Nuclear Regulatory Research

RESEARCH INFORMATION LETTER No. 1Q2 , "STRUCTURAL BUILDING RESPONSE REVIEW: PHASE I OF PROJECT IV OF THE SEISMIC SAFETY MARGINS RESEARCH PROGRAM"

This Research Information Letter (RIL) describes the resultsof a study to review the state-of-the-art in nuclear power plant structural building response computation. This study, initiated under the Seismic Safety Margins Research Program (SSMRP), is not intended to advance the art; rather, it will be used to identify analytical methods for realistically characterizing the seismic response of nuclear power plant structures. The findings of the subject study, as discussed in this RIL, should provide the staff with a basis for detailed review in the areas of analytical methods, structural modeling, uncertainty, nonlinear behavior, methods to account for interactions and nonseismic response. Because this is a RIL for the SSMRP, background information on the SSMRP is provided in addition to a descriptive summary of the structural building response review study.

1.0 Introduction

NRC has established regulations, guides, and licensing review procedures that define seismic safety criteria for nuclear power plant design. These criteria collectively constitute a seismic methodology chain. The seismic safety criteria for nuclear power plant design were develop~d to ensure structural as well as functional safety of buildings and equip-ment supported by buildings, and they depart from the conventional earthquake engineering practice in detail and complexity. The seismic methodology chain is considered sufficiently conservative to ensure safety; however, it is necessary to characterize the overall seismic safety and to improve it by establishing new criteria as may be required.

r:- \~':. ~- ::~~. -::. ·.r-..

Fl ; ,;..·

~. ·, :~~/:

-~~ ·.·j:;;: ~4

f',. '"f,-~~

·~~'

e Harold R. Denton Robert B. Minogue 2

1.1 Seismic Safety Margins Research Program

The SSMRP is developing probabilistic methods that realistically estimate the behavior of nuclear power plants during earthquakes. These probabilistic methods stand in contrast to the conservative methods appropriate to existing seismic design methodology, in which each element of the seismic methodology chain is addressed independently. Since there is uncertainty in each element, conservative assumptions are usually made, and the final result is a summation of several worst-case scenarios. For example, the strongest plausible earthquake is presumed to occur and produce the largest ground motion at the free field of the site. This motion is coupled with the bedrock and the building foundation to produce the worst possible forces and stresses. Such responses are compared to conservative estimates of the ·fragility of each structure or component to determine its survivability. In such a design process, the real safety issue of potential radioactive· release is rarely addressed in the context of a system's assessment.

The objectives of the SSMRP are to develop an improved seismic safety design methodology and to ·develop a methodology to perform earthquake risk assessments of nuclear facilities. Risk will be measured by various failure probabilities and by the probability of release of radioactive materials. The SSMRP approach integrates the elements of the seismic chain, including:

Earthquake characterization

Soil-structure coupling

Structural building response

Subsystem structural response

Local failure

Systematics of how local failures could combine and lead to a release.

Each element will be characterized realistically and probabilistically, rather than conservatively and deterministically. Significant advances in technology will be required to meet the objectives. A multiphase program is underway consisting of eight projects which comprise the program. One of these projects is the Structural Building Response Project, the subject of this memorandum.

1.2 Structural Building Response Project

This project deals with the methodology to be used in the SSMRP for structural building response. The final goal is to determine structural response using state-of-the-art analysis techniques. Structural response serves two main purposes: (1) to develop input motion for the subsequent

... ~ " ··:: .. ,..: ..

··~--~ . . ;.,'f •. .~ .


subsystem analysis and (2) to develop response for estimating structural failure. The models and methods used to determine structural response will be subject to many variables, including:

Structural material properties (modulus .of elasticity, etc.)

Structural dynamic behavior (damping, nonlinear behavior, etc.)

Configuration

Idealization and modeling techniques

Methods of solution

Computer programs

As a first step in the Structural Building Response Project, a review of the state-of-the-art of structural response was performed. This review addressed several major areas of interest:

Structural modeling

Methods of dynamic analysis

Nonlinear behavior of structures and materials

Uncertainty in dynamic structural analysis

These and other issues were addressed in two reports (References 1 and 2). A summary of them_ comprises the remainder of this memorandum.

2.0 Descriptive Summary

This summary is intended to provide a brief description of the contents of the two reports. The reader should consult the reports for further details on the topics presented below.

2.1 Uncertainty

Two inherently different types of uncertainty were identified in the reports: (1) random variability, which is associated with such sta-tistical variations as the natural heterogeneity in material properties; and (2) modeling uncertainty, which is a systematic type of variability related to the limited availability of information, inherent bias in certain models or predictions, consistent errors, or deviations from· reality in material and structural testing.. ·

In fact, few sources of variability can be solely attributed to either random variabi,lity or modeling uncertainty. For example, material properties are certainly a source of random variability; however, the concrete quality control requirements lead to average concrete strengths consistently greater than the nominal values. This latter type of


variability is obviously a modeling or systematic type of uncertainty. Another example of combined random variability and modeling uncertainty is the damping values, which exhibit not only natural irreducible variability, but also a systematic bias in present-day calculations

· because the prescribed values are believed to be less than those experienced in practice, especially at high response levels.

The individual sources of uncertainty were addressed in three broad categories in Reference 2:

Constitutive properties - primarily the elastic constants and strength values for steel, concrete and reinforcing bars, but also the description of the stress-strain behavior over the entire range for use in nonlinear analyses;

Dynamic structural characteristics, include the mass, stiffness and damping characteristics, and the calculated natural frequencies and mode shapes;

Other sources of uncertainty, include modeling techniques, analytical procedures, computer software reliability and effects such as the variation in field construction practices, errors in analysis, design and fabrication and deterioration of members.

Sources of uncertainty are identified as either random variability (RV), modeling uncertainty (MU), or both, and subjective estimates of the uncertainties are provided in a summary (Table 1). This table is based on one that appears in Reference.2. ·

2.2 Nonlinear Behavior

Nonlinear behavior of nuclear power plant structures can result from either geometric or material nonlinearities. However, because of the size and stiffness of these structures, geometric nonlinearities due to large deformations are less likely, and the reports focus on material nonlinearities. Reference 2 discusses nonlinear material characteristics (and the attempts to treat them in analysis with simple idealized force-deformation curves)' in terms of the following assumptions about non-linear behavior:

Lumped plasticity (i.e., ·the formation of a plastic hinge in a frame-type structure when the maximum bending moment reaches the yield moment)

Distributed plasticity (i.e., the formation of plastic regions in a shear wall)

Stiffness degradation in concrete structures due to cracking.

These sources of nonlinearity are treated in dynamic analyses in three different ways:

iii;~- -~.:: . .

~"''·'' f: ·r.

tl r~. :r>-,_. •

!t~~f \" '(;•' i'. ";····

id;: f.:' f~:-.. ~· ,,,s.:.

~ .. '_J~~~

~->-~. ~· . ·~ ;: . ·:~r

. ~: -- .

. "

Harold R. Robert B. • Denton Minogue 5

Detailed multiple-degree-of-freedom (MDOF) inelastic calculations

Single-degree-of-freedom {SDOF) inelastic methods

MDOF elastic analyses

The first technique is the most rigorous, but also the most time-consuming and costly. Nonlinear analysis of MDOF system response to a given time history of ground motion is carried out by step-by-step integration of the equations of motion, dividing the response history into short-time increments and assuming that the properties of the structure remain constant during each increment, but change in ac-cordance with the deformation at the end of each increment. Thus, detailed MDOF nonlinear analysis is actually a sequence of linear analyses of a changing structure.

Because of their simplicity and low cost compared to the MDOF nonlinear analysis technique, SDOF inelastic methods are discussed in the reports. Primary among these methods is the equivalent linear approach to the analysis of simple hysteretic structures for which an equivalent linear system is developed to match the response of the nonlinear system. Reference 2 categorizes the methods for developing the equivalent linear system in terms of three types of input motion - harmonic, random and earthquake loading. The methods discussed in these categories include:

Harmonic Equivalent Linearization

Resonant Amplitude Matching

Dynamic Mass

Constant Critical Damping

Geometric Stiffness

Geometric Energy

Stationary Random Equivalent Linearization

Average Period and Damping

Average Stiffness and Energy

Reference l discusses approximate techniques in terms of the following methods:

Reserve Energy

Inelastic Response Spectrum

Substitute Structure

i· ·. f?' ~- ....


In both reports, the methods were compared to one another not to produce a 11 best 11 met.hod, but rather, to substantiate the claim that approximate methods can be developed for the nonlinear .analysis of nuclear power plant structures (if such an analysis is necessary) to avoid the costly, complex and time-consuming rigorous approach.

2.3 Structural and Component Idealization Methods and Mathematical Models

The reports discuss two major methods of discretization and their applicability to nuclear power plant structures:

Equivalent beam models, in which the mass can be considered as concentrated at a series of points and the stiffness of the overall structure approximates that of a simple cantilever beam, and

Finite element models {both two- and three-dimensional), in which various elements (shells, plates, etc.) describe the overall stiffness.

In general for excitations, axisymmetric shell structures that have a large height-to-radius ratio can be adequately modeled by the equivalent beam method. Typically, chimneys and containment vessels are modeled this way; beam properties are determined from the shell cross section and lumped masses from the dead weight of the shell. Both reports compare this method to the more complex shell (finite element) approach and find good agreement, especially when the lower modes dominate the structural behavior. It was found that for a given structure, the accuracy of the equivalent beam approach is dictated by the total number of masses chosen. Equivalent beam modeling entails two simplifying assumptions:

Plane sections remain plane after deformation

No shape distortion occurs because of the diaphragm action of the floor slabs.

The finite element approach does not require such simplifying assumptions. Various types of elements (e.g., shell, plate, beam and truss elements) describe the overall stiffness. Moreover, local behavior of a structural system can be readily incorporated in a finite element model.

2.4 Analysis Methods

Structural response can be determined by either a time-history or response-spectrum approach, both of which are addressed in the reports.

.:·· .,.'riJ

"

Harold R. Denton Robert B. Minogue 8

2.5 Methods that Account for Interactions

The two reports describe the relationship between structural building response and the coupled soil-structure system. Simplifications of the structural models for soil-structure interaction analysis are discussed.

2.6 Nonseismic Response

The reports briefly discuss the state-of-the-art methods for combining seismic and nonseismic responses.

3.0 Conclusions and Recommendations

Five major areas of seismic response analysis of buildings were reviewed to assess the state-of-the-art.

Findings of this study in the areas of structural modeling, methods of structural analysis, structural damping values, nonlinear behavior of materials and structures, and the uncertainty in structural dynamic analysis will provide the staff with a basis for detailed in-depth review and, in many cases, a confirmation of the technical judgment of the staff. Following are some pertinent conclusions that would be of interest to the staff from the standpoint of licensing review.

a. In the dynamic analysis of structures, the floor slabs which support important systems or components should be modeled properly in the vertical direction of motion. Rigid floor assumption is not sufficient if the floor system is not stiff enough to justify that assumption. Since the floor usually consists of a large number of composite beams and numerous irregularities in floor geometry and thickness, detailed modeling is not usually performed. The floor systems are usually represented by sets of .SDOF systems in the dyn~mic analysis model.

b. A proper distribution of mass and stiffness of the structure is essential in the lumped-mass-beam approach of structural modeling. Many assumptions are generally made by engineers in this approach to simplify the calculation of stiffness characteristics of structures. Close examinations ·of these assumptions are necessary to ensure the reasonableness of the resulting model.

c. Hydrodynamic effects developed during an earthquake cannot be ignored in the-design of power plant structures in cases where the quantity of liquid is.large. Due to the complicated nature of this problem, there is currently no single universally accepted code that can be utilized for computing this effect for generalized conditions.


• d. Decoupling criteria for subsystems currently used in engineering

practice ·employ two numerical ratios; mass ratio Rm' and frequency ratio Rf. In the NRG.Standard Revi.ew Plan, Section 3.7.2, the definitions of these two ratios are not quite clear and may be subject to different interpretations. Reference 2 offers new definitions for these two ratios. These recommendations are one step closer to clear regulation in this area.

e. Under current practice and/or economic reasons, the complicated structural systems are treated by linear seismic analysis. For example, analysis of concrete structures are treated with both gross properties and fully-cracked properties. The cracked properties are usually estimated based on conservative forces developed for the uncracked linear models. It is usually assumed, but not demonstrated, that this analytical approach brackets the true nonlinear response. Considering the inherent difficulties and penalties associated with detailed nonlinear dynamic analysis of MDOF systems, the development of practical, simplified approaches for such nonlinear response calculations is a necessity.

f. Uncertainty of structural characteristics affect virtually every aspect of analytical effort to'predict the actual in-service response of nuclear power plant structures to a strong-motion earthquake. Past studies show that there is large uncertainty in predicting the structural frequencies. Calculated frequencies can deviate substan-tially from the test or observed structural frequencies for a whole range of excitation levels. The results of Reference 2 show that the uncertainty of structural damping values is even greater than that of structural frequencies. NRC Regulatory Guide 1.61 damping values are reasonable for design in view of the current knowledge in this area.

One source of conservatism that has usually gone unnoticed is the current design and analysis practice in the nuclear industry. The member sizes of structures are initially set large enough to enhance nonexceedance in later design modifications. Efforts to. trim down the member sizes are not emphasized since the iterative process of alternating design and analysis is not practical in reality. With so much conservatism built into the design of nuclear

·power plants, the actual stress level under safe shutdown earthquake and operating basis earthquake might be substantially less than the stress levels for which the damping values are originally assigned. Smaller damping values might actually be applicable for final design. Back verification is usually lacking in current practice.

f .. r ) .;.·.:-i ' ~·-(


The results of this state-of-the-art study in the five major areas of seismic response analysis of buildings should be reviewed by the NRR staff for use in the regulatory review process.

Enclosures: 1. Table 1 2. Table 2

cc: F. Schroeder, NRR G. Knighton, NRR S. Chan, NRR W. Anderson, SD

Thomas E. Murley ng Director Office of Nuclear Regulatory Research

e.

REFERENCES

1. Sargent & Lundy Engineers, Structural Building Responses Review Report Prepared for Seismic Safety Margins Research Program, Lawrence Livermore National Laboratory, California, Chicago, IL, Report SL-3759 (Lawrence Livermore National Laboratory, Report UCRL 15183, May 1980). NUREG/CR-1423, Vol. II.

2. J. J. Healey, S. T. Wu and M. Murga, Seismic Safety Margins Research Program (Phase 1) Project IV - Structural Building Response Structural Building Response Review, Ebasco Services, Incorporated, New York City, Lawrence Livermore National Laboratory,· Report UCRL 15185, May 1980. NUREG/CR-1423, Vol. I.

..

; .. -.1 •..

Enclosure l

TABLE 1. Sources, types (random variability, RV, or modeling uncertainty, MU), and estimated magnitudes of uncertainties in the constituitive properties of concrete, concrete reinforcing bars, and structural steel, from Reference 2.

Source of

Ut'lcerU inty

COtiSit~ui:v:: PROB~". tf.S

~

iype

CO!'PRESSIH RV 1 .KUz STRE~UiH

'!~SILE STl!ENGTH

l!Ot:CLUS OF ELASTICITY

CilNCRETi: l"E~ER DlllENS:OH

CC'IC"ETt ~NG B-titS

YIELD STll!FIGTH

R'l ,PU:

RV,11\J

RV,~

•~SIN, i~-•LACE STRENGTH Ts:r· 0.575 f~ • 1100~1.lSf~

f'~ • !l~S!G~ ... ~OMPRESSIVE S1P.:!iG1H

• 'IS?t~s:c~ CYU:4ct~ S":"REUGiH

FOR ~~ < 4000 psi

cov • 1 O': EXCi:LLEllT QC cov • 15~ A~EP..IGE QC COY • zo~ POOR oc FQR f~ >4000 psi

" • 400 osi EXCELLrnr QC ". ~oc PS; AYER.>GE cc a• 900 psi POOR QC

:~ ... ~~c:: snu::1GTM v2 • .,z • o oos•

s":r cy1 • --~~f V • COY

• J!S-;

.. e Enclosure 2

TABLE 2. Sources, types (random variability, RV, or modeling uncertainty, MU), and estimated magnitudes of uncertainties that stem from structural dynamic characteristics and structural modeling, from Reference 2 .

.. ..

so~r~e of i1pe Uncert11nty Es~imate

, Unc:er:dinty

!;''f'N:l. 111 ; C ~EP.IS'iii:S

NATU~AL PERIOD RV,m: ALL BUILDING TYPES

(STEEL,RC & COHPOSl;E) •RATIO Oi' OBSERYEil/COM>UTED PERIOD

SHALL A~PL!TUDE Yl9RATiD~S MEAM RATIO • 0.84) COY • Jl.1~ LARGE AHPLITilE YIBRHIONS MEAN RAo!O • 1.15 COY • ]Q: DISTRIBUTION FUMCT!ON (BOT!' C~SES) LOG NORHAL, G~HHA

•REINFORCED CONCRETE BU!LO!,GS SHALL VIBRATIONS (A"PLI;uDE) ~EAN • A.26: cov • 75-:, LARGE '1 l BR•TIDNS ( A~PL !TUOE) "EA~ • 6.63~ cov • 64:

•STEEL BUiLO:~G~ S~ALL AHPLITUOE Yl9RATICNS MEAN • 1,68: co·1 • 6S~ LARGE •MPL!TUDE VIBRATIONS "EAN • 5.65~ cov • 45: D!STR!aUTIDN FUNCTION (B07k C-S!S) LOG-NORHAL, GAHHA

eCOMPOStiE BUIL!l!NGS SHALL A~PL!TUD[ Yl3~ATIONS HEAN • 2.72: tO't' • 42:

LARGE AMPLITUDE VIBRAT!OMS HEAN • 3. 23; COY • 54: DISTRIBUTION FU~CTION LOG-NORMAL, GAHHA

RV,~ REACTOR BUILDING COMPLEX (SUBJECTIVE ESTIMATE)

•CONCRETE HEAN • 5: RANGE• J: to 15: OlSTRl5UTiOH: UNIFORM

•STEEL HEAN • 2:, RANGE 0.05: - 4:

RV,!'tl •REINFORCED-CONCRETE CONTAINMEN; (SUBJECTIVE ESTIMATE) MEAN • 5: RANGE • 2: to 10: DISTRIBUTION: FIGURE 6,6

RV,~ •REINFORCED CONCRETE BUILDING MEAN • 5.4: AA~GE • I: to 11:

•VARIOUS BUILDINGS, LOW AHPLITCOE MEAN • 3.1: RANGE • D: to 13:

•HIGH RISE BUILDINGS SHEAR WALL TYPE, MEAN• 2.3•: FRAHE TYPE, HEAN • J.4B: CONSERVATISM IN NOMINAL VS ACTU~L DAMPING VALUES • 20-40: cov • 20:

Source of

Uncertainty Type

STRUCTURAL

~ MODEL! NG RY,f'IJ

RV,HU

R"l,l'\I

MASS PROPERTIES AV,PIU

NUMERICAL ACCURACY l!U

Uncert11nty E5tt•ate

•SIMPLE STRUCTURE EXPERIENCED ENGR: co• •2.5~ IUEXPERIENCE~ ENGR: co• • s:

•COMPLEX STRUCTURE EXPERIENCED ENGR: co• • 10: INEXPERIENCED ENGR:cov • 30~ SUBJECTIVE ESTl"ATE cov• 15: SUBJECTIVE ESTIMATE

U.NGE OF UNCERTAINTY • * 201

SUBJECTIVE ESTIMATE RANGE OF UNCERTAINTY •"'- 1 o:

SUBJECTIVE EST:HATE RANGE OF UNCER;A!HTY • ± 5":

f t.

Harold R. Denton Robert B. Minogue

2.4.3 Random Vibration Techniques

8 ··--

Reference 1 discusses 1 third"approach to dynamic seismic analysis, the random vibration method. Sometimes referred to as the power spectral density method, thf s statistical analysis technique uses an ensemble of possible ground-motion histories in contrast to the time-history method, which uses a detenninistic time function, and the response-spectrum technique, which uses a set o{ Slll)()thed response _spectra.

2.5 Met ods that Account for Interactions

The two rep rts describe the relationship between structural buf ldin response and he coupled sotl~structure system. Simplifications the structural ls for soil-stJ:ucture fnteract~on analysis are cussed.

The reports brf efly scuss the state-of-the-art sefsmic·and nonseismic


Distribution chron circ subject Burger cy Burger rf Bagchi Kenne ly cy Sha cy M ley cy arkins cy BMM~~X Budnitz cy

•'

REC D NOTE: A meeting has been held with e User Office and it was agreed that a IL should be prepared to sumnarize and tran mit these reports.

SU:::::: --:~~!~ ·--~;~~- ~~~;~;~~~~- --~~;~~~---··-. -~~!~~~~-~---· ;:!:~~----· ------------------------- ----------------------- ------------------------- ________ ,,, ____ .................... ----- --------------------------- ----------------------

DATE .... ----~J.:J_l~Q ___________ -----~L _l~Q ________ --------~L ___ l~Q _____ -------~l_ ___ l8Q __________ 5_l_ ___ l'QQ ________ 5.L ___ l'QQ ________ _


2.4.3 Random Vibration Techniques

Reference 1 discusses a third:approach to dynamic seismic analysis, the random vibration method. Sometimes referred to as the power spectral density method, this statistical analysis technique uses an ensemble of possible ground-motion histor,es in contras~ to the time-history method, which use a detenni ni st1c time function, and the response-spectrum technique, ich uses a set o~ smoothed response spectra.

2.5 Methods t Account for~Interactions

The two reports cribe the relationship between stnactural bu11df response and the co led soil-structure system. S1mp11ficat1ons structural models fo so11-st~ucture interaction analysis are

2.6 Nonse1sm1c Respons

The reports briefly discus the state-of-the-art met for combining seismic and nonsef smf c respo es.

3.0 Conclusions & Recomnendat1 s:

Six major areas of seismic respons analy s of buildings wre reviewed to assess the state-of-~art.

Findings of this study in the 'areas influence of cutoff frequences on il-s ructure interaction, analysis methods of generation of floo~ ponse s ctra, structural damping

·values, combination of loads, d uncertain fn structural analysis will provide the staff with basis for· deta Jed 1ndepth review and fn many cases, a confinnation f .. the technical j gment of the staff.

If there are any questf - s c:Oricern1ng thfs RIL. C. W. Burger of my s f. ·

Enclosu : · 1. Ta e 1 2. Table 2

·Robert J. Budnitz. Directo ~Office of Nuclear Regulatory

-· ~ RECORD NOTE: ·A meeting has been held ith the User Office and it was agreed that a RIL should be epared to surrunarize and transmit these ~M~MlK reports.

cc: Anderson/SD ChanL/NRR

OFFICE ... GRSR:SRB GRSR:SRB:C GRSR:AD RES:DIR RES:PCB RES:DIR ----------------- - _, -------------------------- ---------------------------- -------------------------- ------------------------- -----------------------Burger/ c Bagchi enneal ly/Shao Murley Larkins · Budnitz

SURNAME ... ---------------------------- -------------------------- --------------------------- ------------------------- ---------------------- ---------·--------- _ ----

DATE ... _51 .LBD------------- _5 __ _ -BD------------ _____ 5 ____ BQ________ _ _____ 5 ____ gc_______ ----~5 ---- BD-------- ___ 5 ---- .so ______ _

Harold R. Denton Robert B. Minogue

--9

d. Decoupling criteria for subsystens currently used fn engineering practice employ two numerical ratios; mass ratio R , and frequency ratio R • Ir. the t~RC Standard Review Plan, Sectfo~ 3. 7 .2, the definit1ons of .thest: biO ratios are not quite clear and n.:,y be s ject to different interpretations. Reference 2 new definitions fo these two ratios. These recomrrendations are one step closer to c ea~ regulation 1n this area.

e •. Under rrent practice and/or economic reasons, the complicated structu -1 syster.:s are treated by linear seismic analysis. For example. alysfs of concrete structures are treated with both gross properties nd fully-cracked properties. The cracked properties are usually ti~ated based on conservative forces deve ed for the uncracked inear models. It ts usually .assumed not demonstrated th t this analytical approach bracket the true nonlinear respon • Considering the fnherent d icult1es and penalties assoc1at d w1th detailed nonlinear nam1c analysis of ~wlt1-degree-of- eedom syste~s. the de opment of practical, simplified approaches for such nonlinear esponse calculations is a necessity.

f. Uncertainty of structural haracte sties affect yirtUally every aspect of analytical effor to p edict the actual in-service response of nuclear power plant struc es to a strong-motion earthquake. Past studies show that ther s large uncertainty in predicting the structural frequencies. cu ted frequencies can deviate substan-tially fro~ the test or serve structural frequencies for a whole range of excitation le ls. The sults of Reference Z that the uncertainty of struc ral damping alues 1s even greater than that of structural freq ~nces. NRC Regu tory Guide 1.61 daMping values are reaso le for design in ew of the current knowledge 1n this area.

One source tonservatism that ha~ usua ly gone unnoticed is the cu ent design and analys-is practi tn nuclear industry. The mem r sizes of structures are tn1ttal set large enough to enhanc nonexceedance 1n later design modif1 at1ons. Efforts to t m down the member s1zes are not er:iphas1 ed ~ince the iterative p ess of alternating design and analysis ts t practical in r. ality. With so DJch conservatism built into e design of nuclear

wer plants, the actual stress level under safe hutdown earthquake and operating basis earthquake might be substantia ly less than the stress levels for which the damping values are originally assigned. Smaller damping values might actually be applicable for final des1gn. Back verification is usually lacking in current practice. .. . .

Robert J. Budnttz, Director Office of t~ucl ear Regulatory . Research

;;o l"T1 ("") -- 0 :;:o l:J

z 0 --i l"T1 ..

(/) )> :::r 03 c: CD __,CD c.. rt ...... t:r :::::l CD l.O

"'C :::r -s ll> ct> (/)

"'C ll> C"" -s ct> ro ro c.. ::I C'+ :::>" 0 ct> __, en c.. c:

~~--s :::r ...... NC'+ CD :::r

ct> ll> ::I c:: c.. Vl

CD M- -s -s ll> 0 ::I ...... en -+i 3 ...... ...... (") rt CD

C"+ ll> :::r ::I CD C.. VI CD -'•

M--s CD :E

"C ll> 0 VI -s M- ll> en l.O . -s

CD CD

llRil c.. l5t: BM-

"18: :::r M ll> C8:: M-!>«: ea: ll> l)Q:: >< :;:o ::is: ..... l:A'J r-~ >< l)Q::

'-= 151::: l:s:: ia: -

User f


• reality. With .so much conservatism built into the design of nuclear power plants, the actual stress level under safe shutdown earthquake: and operating basis earthquake might be substantially less than the stress level for which the damping values are originally assigned. Smaller dampin values might actually be required. ·Back verification is usually lacki in current practice.

g. Topics in structura analysis methods and numerical integration schemes are theoreti 1 in nature, and generally are well developed and well understood. hese methodologies are already programmed into various computer c des. The objective for the structural building response revie in these areas is to present the · or-rnation in an orderly and oncise form for ease of refer e and comparision. Topics in so·l-structure interaction a ysis, sub-system analysis and load co bination are not elab ted on in this review since these topics w1 1 be addressed in rious other SSMRP projects.

Enclosures: 1. .Table 1 2. Table 2

cc: R. Mattson, NRR G. Knighton, \·J. Anderson,

D. t .s.tChan, NR is r1ou ion

chron circ K~MXsub Burger Burger: rf Bag i cy K neally cy

ao cy Murley cy Larkins cy Budnitz cy

•

Research

been held with the User Office ed that a RIL should be prepared

transmit these reports.

Larkins

i

! I

. !

. I

L· .. •.

·~· .--! · ' .

'.

i "t': . . • ....... r,

·.:i-.. ·•:.:

. .'r .. ;

-.t. 't· ... -.

. . ~ \ ·o1:-_ •

. 9· .. .-t

. "·"'

r.··~

;-... . ,~--~\ -.~~-

~- .· ; ,-.

• Harold R. Denton Robert B. Minogue

• 10

The results of this state-of-the-art study 1n the five major areas of seisr~ic respons.e anal.)isi5 of buildings should be reviewed by the NRfl. staff ~or use in the regulatory review process.


cc: F. Schroeder. NRR G. Knighton, NRR S. Chan, NRR W. Anderson. SD

"Ori.;inal signed by Thomas E. Murley, Acting Director .Office of Nuclear Regulatory Research

Record Note: A meeting has been held with the User Office and it was agreed that a RIL Distribution

chron circ

be prepared to surrnnarize and transmit these reports. User Request No. NRR-76-8 .

subject Burger cy Burger rf Bagchi cy Kenneally cy Shao cy Murley cy Larkins cy Murley cy

*see previous concurrence sheets

*GRSR :TA :AD Kenneally

OFFICE ... ~~-~~.R:S_~-------

SURNAME. Burger/cb -------------------

DATE• -------------------NRC FORM 318 (6-77)

,

_!GRSR:S.RB.:.C ----~G.RS.R~J\o ___ . _____ RS,e..:J)J.a ____

---~.?:.~~-~] _______ ---~!l~P---------~ ___ f!1.u.rJ_ey _______ .

-----------------· ------------------· -------------------'ia' U.S. GOVERNMENT PRINTING OFFICE 1977-237-025

~E.S.;..P-CB---- --~T-:-9 ----~a.r. k..iDS.-- ---- v.~~.ey

-----~.hf·_ ------------------

IR

research information letter 102, 'structural building ...seismic safety criteria for nuclear power...

Documents